The testing of semiconductor devices, such as dynamic random access memories, is a very important step in the manufacture of quality semiconductor devices. A number of tests are generally performed on semiconductor devices using automatic test equipment to identify semiconductor devices that are malfunctioning or likely to malfunction in the future. As is well known in the art, although semiconductor devices are very reliable, if they are to fail at all, they tend to fail during the early part of their service life. This early failure phenomenon, known as "infant mortality," provides an opportunity to ensure that semiconductor devices sold by a manufacturer will have a long, useful life.
Although semiconductor devices can be tested at room temperature to identify devices that are malfunctioning or likely to malfunction, such testing is preferably done at other than room temperature. Generally, such environmental testing is at an elevated temperature in "burn-in ovens" but may also be done at a reduced temperature in cooling chambers. The advantage of environmental testing is that the devices are tested in "real world" conditions. As a result, it is possible to identify semiconductor devices that will work properly at room temperature, but will malfunction or fail to meet specifications at the reduced or elevated temperatures that will be encountered during actual use. Further, operating semiconductor devices at an elevated temperature during "burn-in" testing tends to accelerate infant mortality thus allowing quicker identification of semiconductor devices that are likely to malfunction early in their lifetimes.
Post-production environmental testing is generally limited to testing at elevated temperatures, although certain semiconductors, such as those conforming to military specifications, are also tested at reduced temperatures. However, in the interest of brevity, the discussion of environmental testing will be limited to testing at elevated temperatures, it being understood that the discussion also applies to testing at reduced temperatures. Accordingly, environmental testing of semiconductor devices is generally accomplished by placing the semiconductor devices in large burn-in ovens which may hold on the order of 8,000 semiconductor devices. The semiconductor devices are generally temporarily placed on large printed circuit-boards by inserting the semiconductor devices into respective sockets. These printed circuit-boards, known as "burn-in boards" generally have an edge connector that is adapted to plug into a respective connector in a burn-in oven.
A typical bum-board 10 is illustrated in FIGS. 1 and 2. The burn-board 10 has a conventional printed circuit substrate 12 on which an edge connector 14 is formed. A plurality of semiconductor device sockets, indicated generally at 16, are mounted on the substrate 12. In the example shown in FIGS. 1 and 2, 256 semiconductor device sockets 16 are mounted on the substrate 12 in 16 rows and 16 columns. However, it will be understood that greater or fewer semiconductor device sockets 16 may be mounted on the substrate 12 and in a different row and column configuration. A semiconductor device, generally indicated at 18, is inserted in each of the sockets 16. However, the semiconductor devices 18 may be tested without inserting them in sockets 16 by permanently mounting the devices 18 on a substrate 12 to form a functioning circuit.
A typical burn-in oven used in production testing of semiconductor devices is illustrated in FIG. 3. The burn-in oven 20 has an enclosed chamber 22 open at the front of the oven 20 which is normally closed by doors (not shown) pivotally mounted on the oven 20. A plurality of vertically spaced racks 24 are mounted on each side of the chamber 22. The burn-in boards 10 containing the semiconductor devices 18 are positioned on respective racks 24 and inserted into the chamber 22 until the board connector 14 mates with a respective test connector 28 positioned at the rear of the chamber 22. The test connector 28 is coupled to conventional test equipment 30 that performs appropriate tests on the semiconductor devices 18 depending upon the nature of such devices and the degree of testing desired.
Specifications for environmentally testing semiconductor devices generally specify that the tests occur at a specific temperature or narrow range of temperatures. It is therefore desirable to maintain the temperatures in the chamber 22 as uniform as possible throughout the chamber 22. If the chamber 22 was simply heated, the upper portion of the chamber 22 would be hotter than the lower portion since hot air rises. For this reason, burn-in ovens 20 produce an airflow through the chamber 22. By directing air from the same source through both the upper and lower portions of the chamber 22, the upper and lower portions of the chamber 22 can be maintained at approximately the same temperature.
Although vertical temperature gradients can be minimized by air flowing through the chamber 22, doing so does not entirely solve the problem of uneven temperatures of the semiconductor devices 18. The problem is primarily an uneven temperature distribution among the semiconductor devices 18 on each burn-in board 10. In particular, if the airflow across the surface of the burn-in board 10 is not uniform, the temperature of semiconductor devices 18 on the substrate 12 will also not be uniform. Similarly, if the airflow across one of the burn-in boards 10 is different from the airflow across a different burn-board 10, the temperatures of the semiconductor devices 18 on the two boards will differ significantly from each other.
In an attempt to provide uniform temperatures of semiconductor devices 18 on burn-in boards 10, a great deal of effort has been devoted to making the airflow through the chambers 22 of burn-in ovens 20 as uniform as possible. Thus, for example, a large number of relatively small, uniformly distributed vents 34 have been formed on opposite sides of the chamber 22 in an attempt to provide a uniform airflow throughout the entire height, width and length of the chamber 22. However, these attempts have not been entirely successful because, for example, the presence of the racks 24 and the burn-in boards 10 themselves can prevent a uniform airflow even if small vents 34 are uniformly distributed. Furthermore, even if the disruptions caused by the racks 24 and the boards 10 could be eliminated, the airflow would still be somewhat non-uniform because of the effects of the top, bottom, front and back walls of the chamber 22. Thus, despite diligent efforts to improve the uniformity of flow through the chambers 22 of burn-in ovens 20, the temperature of semiconductor devices 18 undergoing burn-in testing may vary excessively. There is therefore a need for an environmental testing device that ensure a greater temperature uniformity of semiconductor devices being environmentally tested.